TechPrufer
Get started
Stripe
Stripe logoStripe · Feature trail

Multi-Region API Call Load Balancing

2 posts · Feb 2025 – Aug 2026

What this trail covers

This post continues the evolution of Stripe's efforts to build and optimize systems for handling API calls across multiple AWS regions. It details how to track customer spend in an omnichannel or multiprocessor environment, addressing the challenges of fragmented payment systems and exploring solutions like the Payment Account Reference (PAR) to gain a unified view. This builds upon previous work in load balancing and resilience.

All Stripe trails
  1. Building a data plane from scratch: Stripe’s own high-performance distributed proxyThis post details the development of Stripe's own high-performance distributed proxy, a data plane built from scratch. It addresses the technical challenge of achieving 99.9995% reliability for API calls processing a significant portion of global GDP. The proxy is designed to minimize failed payment requests by ensuring robust and efficient routing of API traffic across multiple regions.
  2. Load balancing Stripe API calls from multiple AWS regionsThis post details the development of a new high-performance distributed system for load balancing Stripe API calls across multiple AWS regions. It covers architecture design, challenges, and AWS service integration for optimal global performance and rate limit management, specifically focusing on resilient webhook handling and minimizing outages.
  3. Tracking customer spend in an omnichannel or multiprocessor environmentThis post introduces the concept of tracking customer spend in an omnichannel or multiprocessor environment. It highlights the challenges of fragmented payment systems and proposes solutions like the Payment Account Reference (PAR) to achieve a unified view of customer spending. This is a new capability within the broader context of multi-region API call load balancing, focusing on data aggregation and unified customer views across diverse payment processors.